News
Guide
Is AI Making Your Organization More Productive or More Dependent?
Key Takeaways
- OpenAI has shut down Atlas and moved its browsing features into ChatGPT and Chrome.
- The new tools support page interaction, account logins, downloads and autonomous web tasks.
- Enterprise adoption will hinge on security, data controls and oversight of agent actions.
As of July 9, OpenAI has shut down Atlas, the AI-powered browser it launched in October with ChatGPT at its core.
Rather than abandon AI-assisted browsing, the company is redistributing Atlas's agentic AI features across ChatGPT's desktop app and a new Google Chrome extension.
OpenAI Is Abandoning the Standalone Browser
The decision followed an internal directive from Fidji Simo, OpenAI's former CEO of applications, to cut back on "side quests" — a push that also led to the shutdown of AI video-generation tool Sora.
According to OpenAI, the browser is a feature, not a destination, and its capabilities belong in platforms where users already work.
The move comes amid an industry-wide race to embed AI agents into web browsing. Perplexity launched Comet, The Browser Company launched Dia and Google and Microsoft updated Chrome and Edge with AI features.
OpenAI's Chrome extension competes directly with Google's Gemini Side Panel, which performs several of the same tasks.
How OpenAI Is Redistributing Atlas’ Features
Rather than replacing Atlas with a single product, OpenAI is distributing its AI browsing capabilities across multiple surfaces that work together as an integrated workspace.
| Atlas Capability's New Home | What It Adds |
|---|---|
| ChatGPT Chrome extension | AI assistance inside Chrome while browsing. |
| Enhanced ChatGPT desktop browser | Interactive browsing with support for logins, downloads and web interactions. |
| Remote cloud browser | Autonomous browser sessions that allow agents to complete tasks remotely. |
Distribution May Matter More Than the Browser Itself
Moving Atlas features into ChatGPT and Chrome gives OpenAI access to users inside products they already open throughout the workday.
The result is less friction for those who want to adopt AI-assisted browsing, as well as more usage for OpenAI, with users no longer having to download the browser, learn how to use it and persuade IT to approve it.
The strategy also comes with trade-offs. A Chrome extension gives OpenAI reach, but it leaves the company dependent on a browser controlled by Google, one of its largest AI competitors.
Enterprise adoption will also depend on how much control organizations receive over page access, account credentials, data retention and agent actions, especially when the tools can log in, download files or complete tasks without constant user supervision.
Editor's Note: If you’re interested in more OpenAI news…
- OpenAI Puts Codex Coding Agent Inside ChatGPT — The updated agent turns ChatGPT into a hub for autonomous, multi-agent software development.
- OpenAI Releases GPT-5.6 Globally After US Approval — Federal agencies have cleared all three GPT-5.6 variants for worldwide release.
- OpenAI, Broadcom Launch Jalapeño LLM Inference Chip — The new chip is the first in a multi-generation compute platform.